A failed or stalled RAID rebuild is the single most common cause of permanent data loss we see. The instinct to swap a disk and rebuild again is exactly what destroys the array — here's what to do instead.
A rebuild onto a failing disk, or with the disks in the wrong order, can overwrite the data for good. Stop, power down, and don't change anything until it's assessed.
A rebuild reconstructs a failed disk from the others in the array. The problem is that it leans hard on the remaining disks — usually the same age and wear as the one that just failed. If a second disk fails mid-rebuild (common in RAID 5), or the rebuild runs with the disks in the wrong order or the wrong parameters, it can write bad parity across the array and overwrite the real data. That's how a recoverable one-disk fault becomes total, permanent loss — and it's the single most common cause of disaster we see.
The instinct to swap a disk and try again is exactly what to resist.
Turn the array off. Leaving it trying to rebuild onto a failing disk only deepens the damage.
Don't swap more disks, reinitialise, or let the NAS auto-rebuild. Don't run 'repair' tools on the volume.
Before removing anything, label each disk with its slot (drive 1, drive 2, and so on). That order is critical.
Send the disks only — not the enclosure or controller — to a specialist for assessment.
We never rebuild on your live disks. Each member disk is imaged individually and read-only first, so nothing is written to your data. We then reconstruct the array virtually in software — working out the true disk order, RAID level and parity from the images rather than the failing hardware — and recover your files from that. Because your disks are only ever read, the data is preserved while the array is rebuilt offline. We handle every RAID level and the major NAS units (Synology, QNAP, WD and more); you just send the labelled drives.
Quick answers to what people ask most.
Not if the data matters. A rebuild stresses the remaining disks and, if another fails or the configuration is wrong, can overwrite your data permanently. Power it down and have it assessed before rebuilding — a failed rebuild is the top cause of total loss.
Just the drives. Take them out of the enclosure, label each with its bay order (drive 1, drive 2, and so on), and send the disks only — not the NAS box or controller. The bay order matters for the rebuild.
Often, yes. Even with two disks down, we image what we can from each and reconstruct the array virtually. Success depends on the condition of the disks, which is why a diagnostic comes first — and why you shouldn't keep trying to rebuild it.
Yes — Synology, QNAP, WD and others, across Btrfs, ext4, ZFS and SHR. Send the drives labelled with their order and we'll reconstruct the volume; you don't need to send the enclosure.
Power it down, label the disks, and talk to us first. We reconstruct arrays from images without writing to your disks — a free assessment tells you what's recoverable.